CN114046888A - Beam synthesis push-broom radiometer calibration method based on convolutional neural network - Google Patents

Beam synthesis push-broom radiometer calibration method based on convolutional neural network Download PDF

Info

Publication number
CN114046888A
CN114046888A CN202111274687.4A CN202111274687A CN114046888A CN 114046888 A CN114046888 A CN 114046888A CN 202111274687 A CN202111274687 A CN 202111274687A CN 114046888 A CN114046888 A CN 114046888A
Authority
CN
China
Prior art keywords
receiving
push
broom
radiometer
neural network
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202111274687.4A
Other languages
Chinese (zh)
Other versions
CN114046888B (en
Inventor
李�浩
杨小娇
刘淑波
李一楠
宋广南
马严
袁启刚
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Xian Institute of Space Radio Technology
Original Assignee
Xian Institute of Space Radio Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Xian Institute of Space Radio Technology filed Critical Xian Institute of Space Radio Technology
Priority to CN202111274687.4A priority Critical patent/CN114046888B/en
Publication of CN114046888A publication Critical patent/CN114046888A/en
Application granted granted Critical
Publication of CN114046888B publication Critical patent/CN114046888B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01J—MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY
    • G01J5/00—Radiation pyrometry, e.g. infrared or optical thermometry
    • G01J5/52—Radiation pyrometry, e.g. infrared or optical thermometry using comparison with reference sources, e.g. disappearing-filament pyrometer
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Spectroscopy & Molecular Physics (AREA)
  • Radiation Pyrometers (AREA)

Abstract

本发明涉及基于卷积神经网络的波束合成推扫辐射计定标方法，该方法基于卷积神经网络的全链路定标方法，结合其他步骤得到的先验信息，再根据推扫辐射计周期性的对已知微波辐射亮温信息的定标场进行观测，利用定标场的亮温信息以及推扫辐射计系统输出的功率信号，对卷积神经网络的模型参数进行循环、迭代，不断训练，直至搜索得到满足误差阈值时的最优模型参数。推扫辐射计工作时，通过对微波辐射亮温已知的定标场的观测，获取定标数据样本，可以周期性的进行卷积神经网格模型参数训练。

Figure 202111274687

The present invention relates to a beamforming push-broom radiometer calibration method based on convolutional neural network. The calibration field of the known microwave radiation brightness temperature information is randomly observed, and the model parameters of the convolutional neural network are cycled and iterated by using the brightness temperature information of the calibration field and the power signal output by the push-broom radiometer system. Train until the search gets the optimal model parameters that satisfy the error threshold. When the push-broom radiometer is working, the calibration data samples can be obtained by observing the calibration field with known microwave radiation brightness temperature, and the parameters of the convolutional neural grid model can be trained periodically.

Figure 202111274687

Description

Beam synthesis push-broom radiometer calibration method based on convolutional neural network
Technical Field
The invention relates to a beam synthesis push-broom radiometer calibration method based on a convolutional neural network, and belongs to the technical field of space microwave remote sensing.
Background
Different from the systems of the real aperture radiometer and the synthetic aperture radiometer, the beam synthesis push-scan radiometer system carries out beam synthesis in a digital domain, thereby realizing the ultra-high beam efficiency of the antenna electrical property, avoiding the contradiction between mechanical scanning and a large-aperture antenna, and being capable of being used for making up the blank of the near-shore high-precision data. The space-borne application of the beam-forming push-scan radiometer system is still blank, and no open literature is provided for the calibration technology. The acquisition of the weighting coefficients in the beam forming process is directly related to the performance of the push-broom radiometer system after beam forming. The primarily obtained amplitude and phase weighting coefficients are obtained by testing the obtained amplitude and phase inconsistency of a receiving channel of the feed source secondary antenna directional diagram data and the coupling coherent noise, and then the optimal value search is carried out by utilizing a genetic algorithm and a sequence quadratic programming algorithm, however, in the satellite-borne or airborne application of the system, the amplitude and phase characteristics of the feed source secondary antenna directional diagram and the receiving channel can change, and if the initial amplitude and phase weighting coefficient value is used, the performance (mainly comprising main beam efficiency and side lobe performance) of the beam antenna directional diagram after beam synthesis can not meet the index requirement.
Currently, there are three main methods for scaling the amplitude and phase inconsistency among multiple receiving channels: (1) and placing a noise source under the external far-field condition of the radiometer system, performing double-phase correlation operation on output signals of all receiving links to output a complex correlation matrix, and performing least square solution on the correlation matrix by combining the position of the external noise source relative to the radiometer system to obtain the amplitude-phase inconsistency among all receiving channels. (2) The same coherent noise signals are injected into the input ends of the receiving channels, and the two receiving channels are subjected to complex correlation operation and the correlation matrix is subjected to least square solution to obtain the amplitude-phase inconsistency. (3) And injecting a coherent noise signal into a receiving channel, combining an adaptive filter with an LMS algorithm, continuously and circularly iterating by taking the minimum mean square error as a judgment criterion, and updating the amplitude-phase weighting coefficient of beam forming. The method has the following defects: the methods (1) to (3) have high requirements on the power of coherent noise signals, and pairwise multiple correlation of all receiving channels puts extremely high requirements on data processing resources and computing capacity. In order to obtain high main beam efficiency and large-width observation, the number of receiving channels of a beam synthesis push-scanning radiometer system is large, and thousands of receiving channels are usually needed for implementation, and the high requirements of the radiometer system on power consumption, computing resources, storage space and the like face huge challenges, so that the system application is limited.
Disclosure of Invention
The technical problem to be solved by the invention is as follows: the method overcomes the defects of the prior art, provides the beam forming push-broom radiometer calibration method based on the convolutional neural network, improves the detection precision of the beam forming push-broom radiometer system, and provides a method for obtaining high-precision brightness and temperature.
The technical solution of the invention is as follows: a beam synthesis push-broom radiometer calibration method based on a convolutional neural network comprises the following steps:
s1, measuring to obtain a feed secondary antenna directional pattern F of the ring focal reflection surface of the push-broom radiometer combined with all receiving links under the condition of dense feed arrayfeed(i) I is 1-M, and M is the number of receiving links of the push-broom radiometer;
s2, taking each antenna directional diagram after beam synthesis expected to be obtained as an optimization target, and taking a feed source secondary antenna directional diagram F of each receiving link based on push-broom radiometerfeed(i) I is 1 to M, and an optimal beam forming amplitude-phase weighting coefficient matrix C is obtained0(K×N)K is the number of synthesized beams and N is the number of selected receive links in each synthesized beam;
s3, feeding coherent noise signals with the same amplitude and phase to all receiving links of the push-broom radiometer by adopting the coupler to obtain the amplitude and phase inconsistency delta g 'of the receiving channels in the receiving links selected by the wave beam synthesis of the push-broom radiometer'(K×N)；
S4, adopting push-scan radiometer beamsSynthesizing amplitude phase disparity Δ g 'for receive channels in selected receive links'(K×N)Updating the amplitude-phase weighting coefficient C of the beam forming1(K×N)Obtaining a calibrated beam forming antenna directional pattern Fb'eam(k),k＝1～K；
S5, observing the calibration field with known microwave radiation brightness and temperature through the push-scan radiometer system, and observing the power signal P output by the calibration field with known microwave radiation brightness and temperature through the push-scan radiometer systemModelMicrowave radiation brightness and temperature information T as input layer and scaling fieldModelAs an output layer, the scaled beam-forming antenna pattern F obtained in step S4b'eam(k) And K is 1-K and is the initial value of K characteristic graphs of the convolutional neural network model convolutional layer characteristic graph, the convolutional neural network model is trained, the parameters of the convolutional neural network model are determined, the convolutional neural network model equivalent to the push-broom radiometer is obtained, and the calibration of the full link error of the beam synthesis push-broom radiometer is realized.
Preferably, in step S2, a genetic algorithm is first used to initially search for a global optimal solution of the beam-forming amplitude-phase weighting coefficient matrix; then, local search is enhanced by utilizing a sequential quadratic programming algorithm, and finally, an optimal beam forming amplitude-phase weighting coefficient matrix C is obtained0(K×N)。
Preferably, the amplitude-phase inconsistency of the receiving channels in the receiving chain in step S3 includes phase inconsistency and amplitude inconsistency between the receiving channels in the receiving chain.
Preferably, the phase inconsistency between the receiving channels in the receiving chain is obtained by the following method:
and taking one receiving link as a reference link, and performing complex correlation on the voltage signals output by the receiving channels in all the receiving links and the voltage signals output by the receiving channels in the reference link to obtain the phase of the correlation coefficient, namely the phase inconsistency among the receiving channels in the receiving links.
Preferably, the amplitude inconsistency between the receiving channels in the receiving chain is obtained by the following method:
and taking one receiving link as a reference link, and performing autocorrelation on voltage signals output by the receiving channels in all the receiving links to obtain the output power of the receiving channels in the receiving links, wherein the amplitude inconsistency between the receiving channels in the receiving links is obtained by dividing the output power of the receiving channels in each receiving link by the output power of the receiving channels in the reference link.
Preferably, in the step S4, the updated amplitude-phase weighting coefficient matrix C1(K×N)Comprises the following steps:
C1(K×N)＝C0(K×N)·*Δg′(K×N)
where "· denotes the multiplication of corresponding elements of the two matrices, Δ g'(K×N)Each row of elements of (a) corresponds to the amplitude disparity of the receive channels in the N receive chains selected by each beamforming.
Preferably, in the step S6, the convolutional neural network model parameters are optimized by using a gradient optimization method until an error of the convolutional neural network is smaller than a preset threshold.
Compared with the prior art, the invention has the advantages that:
(1) the invention establishes a beam synthesis push-broom radiometer calibration method based on a convolutional neural network, and the amplitude-phase inconsistency of a receiving channel in a receiving link is roughly obtained through periodic coupling coherent noise through the directional diagram data of a secondary antenna of a feed source obtained through ground measurement. Then, observing a calibration field with known microwave radiation brightness and temperature through the push-broom radiometer system, and establishing a reverse model of the push-broom radiometer system by utilizing the multilayer supervision and deep learning characteristics of the convolutional neural network algorithm to finish the whole link error calibration of the beam synthesis push-broom radiometer system;
(2) the coherent noise signals with the same amplitude and phase are injected into the receiving channels of all receiving chains in a coupling mode, the coupling mode has low requirement on noise power, complex correlation is carried out between voltage signals output by the receiving channels in all the receiving chains, self-correlation is carried out on the signals of the receiving channels, the phase and amplitude inconsistency of the receiving channels in the receiving chains is obtained from the correlation coefficient and the power signals, and the uncertainty of the amplitude-phase weighting coefficient of beam forming can be reduced.
(3) The amplitude and phase inconsistency of the receiving channels in all receiving links can be obtained from the correlation coefficient and the power signal output by the radiometer system only by performing one-time complex correlation operation on the output signal of the receiving channel in all receiving links and a reference channel (for example, selecting the first receiving channel) and performing self-correlation operation on the receiving channel, so that the amplitude and phase inconsistency calibration between the receiving channels in the receiving links is preliminarily completed by using few operation times, the calibration result of a subsequent calibration method can be preliminarily restricted, the phenomenon that the subsequent convolutional neural network model search process falls into a local optimal value is avoided, and the search of the optimal parameter of the convolutional neural network is facilitated.
(4) The invention provides a convolutional neural network-based method for calibrating the weighting coefficients of beam forming, overcomes the defect that the traditional multiple receiving links only calibrate the amplitude-phase inconsistency existing in the receiving channel part, and can realize the calibration of the full link error of the push-broom radiometer system.
(5) The invention provides a method for calibrating errors of a plurality of receiving links simultaneously aiming at the problem that the direct output performance of a push-broom radiometer system is reduced due to amplitude-phase inconsistency among a plurality of receiving links, and the method can be applied to spaceborne, airborne and ground multi-beam multi-channel receiving radiometers and radar systems and improves the application performance of the system.
Drawings
FIG. 1 is a schematic diagram of a push-broom radiometer system according to an embodiment of the present invention;
FIG. 2 is a diagram of a beam forming link according to an embodiment of the present invention;
FIG. 3(a) is a schematic diagram of antenna patterns tested according to an embodiment of the present invention;
fig. 3(b) is a schematic diagram of antenna patterns that are desirable for embodiments of the present invention;
FIG. 4 is a flow chart of a method of an embodiment of the present invention;
FIG. 5 is a flow chart of a convolutional neural network according to an embodiment of the present invention;
FIG. 6 is a convolutional neural network model training process according to an embodiment of the present invention.
Detailed Description
The following describes embodiments of the present invention in further detail with reference to the accompanying drawings.
Different from the traditional real-aperture radiometer system, the beam synthesis push-broom radiometer can acquire the microwave radiation brightness temperature of a target scene with high resolution and wide observation field of view without cone scanning, overcomes the engineering realization problem of mechanical scanning of a large-aperture antenna by high resolution, and becomes a research hotspot in the field of microwave remote sensing. However, when the radiometer system is applied, due to the non-idealities of the environment, the system and the like, an error exists in a receiving link (mainly comprising an antenna and a receiving channel) of the system, and at the moment, microwave radiation brightness and temperature information of an observation target scene cannot be correctly solved through a power signal output by the system and information such as a feed source secondary antenna directional diagram and the like acquired by the radiometer system on the ground. In order to obtain high precision brightness temperature information, the error of the push-scan radiometer system must be calibrated. However, the conventional radiometer system has no beam forming step, so the conventional calibration method cannot be applied to error calibration of the beam forming push-broom radiometer system.
The push-broom radiometer observes the earth by means of satellite-borne, and can measure the microwave radiation brightness and temperature of earth target scenes such as atmosphere, ocean, polar regions and the like.
As shown in fig. 1, the beam-forming push-scan radiometer system includes: the device comprises a ring focus reflecting surface, a dense feed source array, M feed sources, M vertical polarization receiving channels, M horizontal polarization receiving channels and a data processor. The following description is given by taking M receiving links as objects, where the receiving link is composed of a feed source and a receiving channel, and the receiving channel is a vertical polarization receiving channel or a horizontal polarization receiving channel.
For each receiving link, the ring focus reflecting surface reflects and focuses microwave radiation brightness and temperature signals of an observation target scene to a feed source of a feed source array, the feed source receives the microwave radiation brightness and temperature signals, converts the microwave radiation brightness and temperature signals into radio frequency signals, and sends the radio frequency signals to a receiving channel to amplify, down-convert and filter the signals into intermediate frequency signals; the voltage signals output by the receiving channels in all the receiving links are sent to the data processor, and the data processor carries out complex correlation or autocorrelation operation on the voltage signals output by the receiving channels in all the receiving links.
As shown in fig. 2, the push-scan radiometer system has M receiving chains, and in order to form beams meeting requirements, each beam is obtained by weighted summation of output signals of receiving channels in the N receiving chains, and the M feed sources can form K beams, and the receiving chains among the K beams have cross multiplexing. The data processor completes beam synthesis and autocorrelation according to the output signals of the receiving channels in the N receiving links selected by each beam to obtain a power signal for outputting each beam, and microwave radiation brightness and temperature information of a target scene is obtained by adopting a two-point calibration method according to the power signal of each beam.
The invention provides a beam forming push-broom radiometer calibration method based on a convolutional neural network, which comprises the following specific steps of:
s1, measuring to obtain a feed secondary antenna directional pattern F of the ring focal reflection surface of the push-broom radiometer combined with all receiving links under the condition of dense feed arrayfeed(i) I is 1-M, and M is the number of receiving links of the push-broom radiometer;
the preferred scheme is as follows:
each feed source secondary antenna directional diagram specifically comprises: placing a ring focal reflecting surface and a dense feed source array of the push-broom radiometer system on a central target point (provided by a measuring field) specified by the spherical near field mechanical arm according to the geometric position center of a feed source array central unit in the spherical near field; the mechanical arm of the spherical near field is rotated and moved by controlling the scanning mode of the motor, so that the radio frequency emission signals (provided by a measurement field) of the spherical near field are positioned at different positions under the array coordinate system of the feed source, and the full space solid angle of the feed source is covered, thus obtaining all the feed sourcesRadio frequency signals at a full spatial solid angle; dividing the radio frequency signals of all the feed sources by the amplitude and phase of the spherical near-field radio frequency transmitting signal (provided by the measuring field) to obtain a secondary antenna directional diagram F of each feed sourcefeed。
S2, taking each antenna directional diagram after beam synthesis expected to be obtained as an optimization target, and taking a feed source secondary antenna directional diagram F of each receiving link based on push-broom radiometerfeed(i) I is 1 to M, and an optimal beam forming amplitude-phase weighting coefficient matrix C is obtained0(K×N)K is the number of synthesized beams and N is the number of selected receive links in each synthesized beam;
the preferred scheme is as follows:
as shown in fig. 3(a) and fig. 3(b), the performance of the secondary antenna pattern of a single feed is difficult to further improve, and in order to obtain an antenna pattern with a narrow beam and low side lobes, a beam is synthesized by a plurality of feeds to obtain an antenna pattern after beam synthesis. The secondary antenna patterns of the feed source numbers 1 and 2 … … M are respectively Ffeed(1)、Ffeed(2)……Ffeed(M) the antenna pattern F corresponding to the synthesized beam numbers 1 and 2 … … Kbeam1、Fbeam2……FbeamK. The beamformed antenna pattern is the desired target (e.g., the beamformed antenna pattern shown in fig. 3 (b)), the feed secondary antenna pattern is obtained in step S1, and the beamformed antenna pattern has the following relationship:
Figure BDA0003329753110000071
in the formula, Ffeed(i) The feed secondary antenna pattern representing the ith receive chain can be abbreviated as matrix multiplication:
Fbeam(K×1)＝C0(K×N)·*F′feed(K×N)
Figure BDA0003329753110000081
in the formula (I), the compound is shown in the specification,". denotes the multiplication of corresponding elements of the two matrices, F'feed(K×N)Each row in the matrix represents a corresponding N feed secondary antenna patterns in the N receive chains selected for each beam.
Solving the beam forming amplitude-phase weighting coefficient matrix C of the above formula0The unknown number is far greater than the equation number, so that the method of directly using matrix inversion to obtain the weighting coefficient matrix is a pathological mathematical process, the obtained solution is unstable, and great errors exist.
The genetic algorithm used in the step is a highly parallel, random and self-adaptive search algorithm developed by taking the natural selection and evolution mechanism of the biology as a reference. Using a group search technology, representing the group as a group of problem solutions, generating a new generation of group by applying a series of genetic operations such as constraint, selection, intersection, variation and the like expressed in the relational expression between the beam-forming antenna directional diagrams on the current group, gradually evolving the group to a state containing an approximately optimal solution, and quickly finding out an overall optimal solution;
and then constructing a quadratic programming subproblem at each iteration point by using a sequential quadratic programming algorithm, taking the solution of the subproblem as an iteration search direction, performing one-dimensional search along the direction, and finally approaching an optimal solution C through repeated iteration0(K×N)。
Wave beam weighting coefficient matrix C obtained by searching through genetic algorithm and sequence quadratic programming algorithm0(K×N)Combining the antenna directional pattern data of the feed source, the actual antenna directional pattern F after each wave beam is synthesized can be obtainedbeam。
Initial value C of amplitude-phase weighting coefficient of beam forming0The antenna directional diagram is obtained based on a feed source antenna directional diagram obtained through ground measurement, and influence brought by changes of amplitude-phase characteristics of a feed source and a receiving channel in a receiving link is not considered. Due to non-idealities in engineering implementation and variations in operating environment temperature, variations in amplitude-phase non-uniformity of the receive channel may result. Meanwhile, the directional diagram of the secondary antenna of the feed source can be changed due to the change of the expansion, the environmental temperature and the like of the ground-developed antenna after the antenna is in orbit, and the actual beam combination is established during the satellite-borne applicationAntenna direction diagram model
Figure BDA0003329753110000082
Comprises the following steps:
Figure BDA0003329753110000091
in the formula, Δ g (i) represents the magnitude-phase characteristic error of the ith reception link, and Δ f (i) represents the secondary antenna pattern error of the ith reception link feed. Therefore, the on-orbit beam synthesis antenna directional pattern has a larger difference from the beam synthesis antenna directional pattern obtained by the ground test.
S3, feeding coherent noise signals with the same amplitude and phase to all receiving links of the push-broom radiometer by adopting the coupler to obtain the amplitude and phase inconsistency delta g 'of the receiving channels in the receiving links selected by the wave beam synthesis of the push-broom radiometer'(K×N)；
The amplitude-phase disparity of the receive channels in the receive chain includes phase disparity and amplitude disparity between the receive channels in the receive chain.
The method comprises the steps of feeding coherent noise signals with the same amplitude and phase into receiving channels in all receiving links through a coupler, taking one receiving link as a reference link (selecting a first receiving channel), performing complex correlation on voltage signals output by the receiving channels in all the receiving links and voltage signals output by the receiving channels in the reference link, and obtaining the phase of a correlation coefficient, namely the phase inconsistency between the receiving channels;
using a certain receiving link as a reference link (selecting a first receiving channel), performing autocorrelation on voltage signals output by receiving channels in all receiving links, and dividing an output power signal by the receiving channels in the reference link to obtain amplitude inconsistency among the receiving channels in the receiving links, and thus obtaining the amplitude inconsistency Δ g of the receiving channels in all receiving links, wherein the preferable scheme is as follows:
as known from the engineering realization of the receiving channel, the amplitude-phase inconsistency characteristic of the receiving channel is large, and in order to obtain the global optimumIt is necessary to feed coherent noise signals of the same amplitude and phase to all reception channels via couplers. Multiplying the voltage signals output by all the receiving channels with the voltage signal output by the reference channel (selecting the first receiving channel) to obtain the real part V of the complex correlation coefficientreal(ii) a All the receiving channel output voltage signals are subjected to 90-degree phase shift and then multiplied by the voltage signals output by the reference channel to obtain an imaginary part V of the cross-correlation coefficientimag(ii) a The phase inconsistency between the receive channels can be solved from the complex correlation coefficients:
Figure BDA0003329753110000101
in the formula, atan represents an arctangent function of a trigonometric function.
The voltage signals output by the receiving channels in all receiving links are multiplied by the signals of the receiving channels to obtain the autocorrelation value of the receiving channels, namely the Power signal Power. Power signal Power of receiving channel in each receiving chainiI 1-M and a reference receiving channel (selecting the first receiving channel) Power signal Power1The ratio of (a) to (b) is the amplitude inconsistency of the receiving channels in the receiving link:
Figure BDA0003329753110000102
in the formula, piThe amplitude disparity of the receive channels in the ith receive chain.
The amplitude-phase inconsistency Δ g of the receive channel is:
Figure BDA0003329753110000103
amplitude-phase inconsistency matrix delta g 'of receive channels in receive chain selected by push-scan radiometer beam synthesis'(K×N)Expressing the following formula, each row in the matrix represents the amplitude phase inconsistency of the receiving channels in the N receiving chains selected by each beam.
Figure BDA0003329753110000104
S4 amplitude phase inconsistency deltag 'of receiving channels in receiving chain selected by adopting beam synthesis of push-scan radiometer'(K×N)Updating the amplitude-phase weighting coefficient C of the beam forming1(K×N)Obtaining a calibrated beam forming antenna directional pattern Fb'eam(k),k＝1～K；
The preferred scheme is as follows:
the beam combination antenna directional diagram after the amplitude-phase inconsistency among the receiving channels in the receiving link is calibrated is as follows:
Figure BDA0003329753110000111
updated amplitude-phase weighting coefficient matrix C1：
C1(K×N)＝C0(K×N)·*Δg′(K×N)
Where "· denotes the multiplication of corresponding elements of the two matrices, Δ g'(K×N)Each row in (a) represents the amplitude phase disparity of the receive channels in the N receive chains selected for each beam. After the amplitude-phase inconsistency of the receiving channel is obtained, an error still exists between the actual antenna directional pattern of the on-track beam synthesis and the beam synthesis directional pattern calibrated in step S4, and the residual error is an error of the on-track feed source antenna directional pattern and a residual error after the amplitude-phase inconsistency of the receiving channel is calibrated. Step S3 carries out preliminary calibration to the beam forming antenna directional diagram, avoids the situation that the partial optimal value falls in the convolution neural network searching process adopted in step S5, and is beneficial to obtaining the result of the full link optimization of the beam forming push-broom radiometer system by the method.
S5, observing the calibration field with known microwave radiation brightness and temperature through the push-scan radiometer system, and observing the power signal P output by the calibration field with known microwave radiation brightness and temperature through the push-scan radiometer systemModelMicrowave radiation brightness and temperature information T as input layer and scaling fieldModelAs an output layer, the scaled beam-forming antenna pattern F obtained in step S4b'eam(k) And K is 1-K and is the initial value of K characteristic graphs of the convolutional neural network model convolutional layer characteristic graph, the convolutional neural network model is trained, the parameters of the convolutional neural network model are determined, the convolutional neural network model equivalent to the push-broom radiometer is obtained, and the calibration of the full link error of the beam synthesis push-broom radiometer is realized.
The preferred scheme is as follows:
as shown in fig. 5, the convolutional neural network is divided into: an input layer, a convolutional layer, a pooling layer, a full-link layer, and an output layer. An input layer: the push-broom radiometer system observes a power signal output by a target scene, namely an autocorrelation value after beam synthesis. The convolution layer is used for feature extraction and is a core for realizing the convolutional neural network, different convolution kernels are used for extracting different features, and the more convolution kernels are, the more features can be extracted from input data. The input of the neuron of each feature extraction layer is connected with the local part of the previous layer, and the feature of the local area is obtained through the neuron. The pooling layer has the functions of reducing the data volume of the convolution layer and improving the operation speed of the convolution neural network on the basis of ensuring the integrity of information. The fully-connected layer is actually part of a hidden layer in the neural network, and the neurons of the fully-connected layer are connected with nodes on the neurons of the pooling layer of the previous layer, but the neurons in the same fully-connected layer are not connected with each other. The output layer is the microwave radiation brightness and temperature information of the observation target scene.
Calculation of convolutional layer:
Figure BDA0003329753110000121
in the formula (I), the compound is shown in the specification,
Figure BDA0003329753110000122
for the j 'th feature map of the l layer, i' represents the number of selected convolution kernel rows, Kil'j'Is the convolution kernel of the l-th layer, f (-) is the excitation function,
Figure BDA0003329753110000123
as a bias parameter, Mj'To select a set of input feature maps. Combining the preliminarily scaled beam-forming antenna directional pattern obtained in step S4 to obtain the initial values of K characteristic patterns of the first layer of the convolutional layer
Figure BDA0003329753110000124
The gradient for the convolutional layer l-1 followed by the connection to the next convolutional layer l is:
Figure BDA0003329753110000125
in the formula (I), the compound is shown in the specification,
Figure BDA0003329753110000126
for the jth feature map of layer l of the convolutional layer and the error signal of layer l-1 of the convolutional layer, up represents the lift sampling operation, and u, v represent the position coordinates of each element of the matrix.
The principle of the pooling layer calculation is that the size of each output feature map is a reduced version of the input feature map, as shown in the following formula:
Figure BDA0003329753110000127
where down (-) is a downsampling function, β is a multiplicative bias parameter, and b is an additive bias parameter.
Gradient calculation of the pooling layer:
Figure BDA0003329753110000131
Figure BDA0003329753110000132
the full-connection layer keeps full connection between neurons of each layer to simulate a convolutional neural network modelType error result oj'：
Figure BDA0003329753110000133
Wherein k 'represents a k' th convolutional layer,
Figure BDA0003329753110000134
error signal for each profile j 'of the k' th layer in the convolutional layer. When o isj'When the expected value is met (the user-defined error threshold value is smaller, the obtained model is more accurate, the calibration effect is better, but the operation time is long due to the increase of the operation amount), the gradient updating of the convolution layer and the pooling layer is stopped, the global optimum value searching is completed, and the returned value
Figure BDA0003329753110000135
The convolution neural network model is suitable for the beam-forming push-broom radiometer system.
As shown in fig. 6, a calibration field (optionally, a cold air calibration field, a sea surface calibration field, a rainforest or a desert calibration field) with known microwave radiation brightness temperature is observed through the push-broom radiometer system, and the microwave radiation brightness temperature information of the calibration field is TModelWhen the output power signal of the push-broom radiometer system is PModel，PModelAs input layer, TModelAs the output layer, beam-synthesized antenna pattern F 'scaled in step S4'beamAs the initial value of the convolutional neural network model. And training the model through continuous circulation and iteration until the optimal model parameters meeting the error threshold are obtained by searching, wherein the optimal model parameters comprise parameter selection of the convolutional layer and the pooling layer.
In summary, the full-link calibration method based on the convolutional neural network of the present invention combines the prior information obtained in other steps, periodically observes the calibration field of the known microwave radiation brightness and temperature information according to the push-broom radiometer, and performs loop and iteration on the model parameters of the convolutional neural network by using the brightness and temperature information of the calibration field and the power signal output by the push-broom radiometer system, and continuously trains until the optimal model parameters satisfying the error threshold are obtained by searching. When the push-broom radiometer works, calibration data samples are obtained through observation of a calibration field with known microwave radiation brightness and temperature, and parameter training of a convolution neural grid model can be periodically carried out.
Based on the optimal convolutional neural network model parameters obtained after calibration, when the optimal convolutional neural network model parameters are applied to a beam forming push-broom radiometer to observe a target scene with unknown brightness temperature of other microwave radiometers, the microwave radiation brightness temperature information with high detection precision can be obtained by combining power signals output by a push-broom radiometer system, and therefore high-precision calibration of full link errors of the beam forming push-broom radiometer is achieved.
Those skilled in the art will appreciate that the details of the invention not described in detail in this specification are well within the skill of those in the art.

Claims (7)

1.基于卷积神经网络的波束合成推扫辐射计定标方法，其特征在于步骤如下：1. The beamforming push-broom radiometer calibration method based on convolutional neural network is characterized in that the steps are as follows: S1、测量得到推扫辐射计的环焦反射面结合密集馈源阵列条件下的所有接收链路的馈源次级天线方向图Ffeed(i),i＝1～M，M为推扫辐射计接收链路数目；S1. Measure and obtain the feed secondary antenna pattern F feed (i) of all receiving links under the condition of the annular focal reflector of the push-broom radiometer combined with the dense feed array, i=1～M, where M is the push-broom radiation count the number of receiving links; S2、以期望得到的各波束合成后天线方向图作为优化目标，基于推扫辐射计各接收链路的馈源次级天线方向图Ffeed(i),i＝1～M，得到最优波束合成幅相加权系数矩阵C0(K×N)，K为合成波束数目，N为每个合成波束中所选择接收链路数目；S2. Take the desired antenna pattern after beam synthesis as the optimization target, and obtain the optimal beam based on the feed secondary antenna pattern F feed (i) of each receiving link of the push-broom radiometer, i=1～M A composite amplitude-phase weighting coefficient matrix C 0 (K×N) , where K is the number of composite beams, and N is the number of selected receive links in each composite beam; S3、采用耦合器对推扫辐射计所有接收链路馈入幅度相位相同的相干噪声信号，得到推扫辐射计各波束合成选择的接收链路中接收通道的幅相不一致性Δg′(K×N)；S3. Use couplers to feed coherent noise signals with the same amplitude and phase to all receiving links of the push-broom radiometer, and obtain the amplitude-phase inconsistency Δg′ (K× N) ; S4、采用推扫辐射计各波束合成选择的接收链路中接收通道的幅相不一致性Δg′(K×N)，更新波束合成的幅相加权系数C1(K×N)，得到定标后的波束合成天线方向图F′beam(k),k＝1～K；S4. Adopt the amplitude-phase inconsistency Δg′ (K×N) of the receiving channel in the receiving chain selected by each beamforming of the push-broom radiometer, update the beamforming amplitude-phase weighting coefficient C1 (K×N) , and obtain the calibration After beam-synthesized antenna pattern F′ beam (k), k=1～K; S5、通过推扫辐射计系统观测微波辐射亮温已知的定标场，以推扫辐射计观测微波辐射亮温已知的定标场输出的功率信号PModel作为输入层、定标场的微波辐射亮温信息TModel作为输出层，步骤S4所得到的定标后的波束合成天线方向图F′beam(k),k＝1～K为卷积神经网络模型卷积层特征图K个特征图的初始值，对卷积神经网络模型进行训练，确定卷积神经网络模型参数，得到与推扫辐射计等效的卷积神经网络模型，实现波束合成推扫辐射计的全链路误差的定标。S5. Observe the calibration field with known microwave radiation brightness temperature through the push-broom radiometer system, and use the push-broom radiometer to observe the power signal P Model output from the calibration field with known microwave radiation brightness temperature as the input layer and the calibration field. The microwave radiation brightness temperature information T Model is used as the output layer, and the calibrated beamforming antenna pattern F′ beam (k) obtained in step S4, k=1～K is the convolutional neural network model convolution layer feature map K The initial value of the feature map, the convolutional neural network model is trained, the parameters of the convolutional neural network model are determined, the convolutional neural network model equivalent to the push-broom radiometer is obtained, and the full-link error of the beamforming push-broom radiometer is realized. calibration. 2.根据权利要求1所述的基于卷积神经网络的波束合成推扫辐射计定标方法，其特征在于所述步骤S2先采用遗传算法初步搜索得到波束合成幅相加权系数矩阵的全局最优解；再利用序列二次规划算法加强局部搜索，最终得到最优波束合成幅相加权系数矩阵C0(K×N)。2. the beamforming push-broom radiometer calibration method based on convolutional neural network according to claim 1, it is characterized in that described step S2 first adopts genetic algorithm preliminary search to obtain the global optimum of beamforming amplitude phase weighting coefficient matrix Then use the sequential quadratic programming algorithm to strengthen the local search, and finally obtain the optimal beamforming amplitude and phase weighting coefficient matrix C 0(K×N) . 3.根据权利要求1所述的基于卷积神经网络的波束合成推扫辐射计定标方法，其特征在于所述步骤S3中接收链路中接收通道的幅相不一致性包括接收链路中接收通道之间的相位不一致性和幅度不一致性。3. the beamforming push-broom radiometer calibration method based on convolutional neural network according to claim 1, is characterized in that the amplitude-phase inconsistency of the receiving channel in the receiving chain in the described step S3 comprises receiving in the receiving chain. Phase inconsistency and amplitude inconsistency between channels. 4.根据权利要求3所述的基于卷积神经网络的波束合成推扫辐射计定标方法，其特征在于所述接收链路中接收通道之间的相位不一致性通过如下方法得到：4. the beamforming push-broom radiometer calibration method based on convolutional neural network according to claim 3, is characterized in that the phase inconsistency between the receiving channels in the described receiving link is obtained by the following method: 以某一个接收链路作为参考链路，将所有接收链路中接收通道输出的电压信号与参考链路中接收通道输出的电压信号作复相关，得到的相关系数的相位即为接收链路中接收通道之间的相位不一致性。Taking a certain receiving link as a reference link, the voltage signals output by the receiving channels in all receiving links are complexly correlated with the voltage signals output by the receiving channels in the reference link, and the phase of the obtained correlation coefficient is the phase in the receiving link. Phase inconsistency between receive channels. 5.根据权利要求3所述的基于卷积神经网络的波束合成推扫辐射计定标方法，其特征在于所述接收链路中接收通道之间的幅度不一致性通过如下方法得到：5. the beamforming push-broom radiometer calibration method based on convolutional neural network according to claim 3, is characterized in that the amplitude inconsistency between the receiving channels in the described receiving link is obtained by the following method: 以某一个接收链路作为参考链路，所有接收链路中接收通道输出的电压信号作自相关得到接收链路中接收通道的输出功率，各接收链路中接收通道输出功率与参考链路中接收通道的输出功率相除，则得到接收链路中接收通道之间的幅度不一致性。Taking a certain receiving chain as the reference chain, the output power of the receiving channel in the receiving chain is obtained by auto-correlation of the voltage signals output by the receiving channels in all receiving chains. Dividing the output power of the receive channels yields the amplitude inconsistency between receive channels in the receive chain. 6.根据权利要求1所述的基于卷积神经网络的波束合成推扫辐射计定标方法，其特征在于所述步骤S4中，更新后的幅相加权系数矩阵C1(K×N)为：6. the beamforming push-broom radiometer calibration method based on convolutional neural network according to claim 1, is characterized in that in described step S4, the updated amplitude phase weighting coefficient matrix C 1 (K × N) is : C1(K×N)＝C0(K×N).*Δg′(K×N) C 1(K×N) =C 0(K×N) .*Δg′ (K×N) 式中，“.*”表示两个矩阵的对应元素相乘，Δg′(K×N)的每一行元素对应每一个波束合成选择的N个接收链路中接收通道的幅相不一致性。In the formula, ".*" represents the multiplication of the corresponding elements of the two matrices, and each row of elements of Δg' (K×N) corresponds to the amplitude and phase inconsistency of the receiving channels in the N receiving chains selected by each beamforming. 7.根据权利要求1所述的基于卷积神经网络的波束合成推扫辐射计定标方法，其特征在于所述步骤S6采用梯度优化方法对卷积神经网络模型参数进行优化，直至卷积神经网络的误差小于预设阈值。7. the beamforming push-broom radiometer calibration method based on convolutional neural network according to claim 1, is characterized in that described step S6 adopts gradient optimization method to optimize convolutional neural network model parameters, until convolutional neural network. The error of the network is less than the preset threshold.
CN202111274687.4A 2021-10-29 2021-10-29 Calibration method of beamforming pushbroom radiometer based on convolutional neural network Active CN114046888B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111274687.4A CN114046888B (en) 2021-10-29 2021-10-29 Calibration method of beamforming pushbroom radiometer based on convolutional neural network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111274687.4A CN114046888B (en) 2021-10-29 2021-10-29 Calibration method of beamforming pushbroom radiometer based on convolutional neural network

Publications (2)

Publication Number Publication Date
CN114046888A true CN114046888A (en) 2022-02-15
CN114046888B CN114046888B (en) 2024-02-20

Family

ID=80206732

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111274687.4A Active CN114046888B (en) 2021-10-29 2021-10-29 Calibration method of beamforming pushbroom radiometer based on convolutional neural network

Country Status (1)

Country Link
CN (1) CN114046888B (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115932843A (en) * 2022-10-28 2023-04-07 西安空间无线电技术研究所 Push-broom radiometer digital beam synthesis method based on multilayer perception network
CN117196418A (en) * 2023-11-08 2023-12-08 江西师范大学 An artificial intelligence-based reading teaching quality assessment method and system

Citations (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2004028731A (en) * 2002-06-25 2004-01-29 Natl Space Development Agency Of Japan Calibration method for total power microwave radiometer for satellite
CN101936778A (en) * 2010-08-05 2011-01-05 中国兵器工业第二○六研究所 Radiometer calibration method
CN101975971A (en) * 2010-09-30 2011-02-16 中国科学院国家天文台 On-star calibration method for lunar microwave sounder
US20130135139A1 (en) * 2011-11-24 2013-05-30 Chun-Yen Huang Calibration devices and calibration methods
CN105988146A (en) * 2015-01-29 2016-10-05 中国科学院空间科学与应用研究中心 Application data processing method of spaceborne microwave radiometer
CN109212496A (en) * 2018-09-26 2019-01-15 西安空间无线电技术研究所 A kind of satellite-borne microwave radiometer antenna error bearing calibration
CN110617889A (en) * 2019-08-28 2019-12-27 西安空间无线电技术研究所 High-stability testing method applied to synthetic aperture microwave radiometer
CN110716185A (en) * 2019-10-22 2020-01-21 国家卫星气象中心 An on-orbit radiation correction method for spaceborne microwave radiometers
CN111538000A (en) * 2020-03-30 2020-08-14 西南电子技术研究所（中国电子科技集团公司第十研究所） Brightness Temperature Inversion Imaging Method for Uniform Circular Array Synthetic Aperture Radiometer
CN111929651A (en) * 2020-07-30 2020-11-13 西安空间无线电技术研究所 Method for calibrating phase residual error of synthetic aperture radiometer system
CN112285659A (en) * 2020-07-30 2021-01-29 西安空间无线电技术研究所 A method for on-orbit update of brightness temperature reconstruction matrix based on synthetic aperture radiometer
DE112019003384T5 (en) * 2018-07-02 2021-03-25 Elecs Industry Co., Ltd. PROCEDURE FOR CALIBRATING A MICROWAVE RADIOMETER
CN113218510A (en) * 2021-05-13 2021-08-06 上海航天测控通信研究所 Feed source mouth surface external calibration device of satellite-borne one-dimensional synthetic aperture radiometer
CN113484917A (en) * 2021-06-30 2021-10-08 中国电波传播研究所（中国电子科技集团公司第二十二研究所） Automatic calibration method for microwave radiometer

Patent Citations (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2004028731A (en) * 2002-06-25 2004-01-29 Natl Space Development Agency Of Japan Calibration method for total power microwave radiometer for satellite
CN101936778A (en) * 2010-08-05 2011-01-05 中国兵器工业第二○六研究所 Radiometer calibration method
CN101975971A (en) * 2010-09-30 2011-02-16 中国科学院国家天文台 On-star calibration method for lunar microwave sounder
US20130135139A1 (en) * 2011-11-24 2013-05-30 Chun-Yen Huang Calibration devices and calibration methods
CN105988146A (en) * 2015-01-29 2016-10-05 中国科学院空间科学与应用研究中心 Application data processing method of spaceborne microwave radiometer
DE112019003384T5 (en) * 2018-07-02 2021-03-25 Elecs Industry Co., Ltd. PROCEDURE FOR CALIBRATING A MICROWAVE RADIOMETER
CN109212496A (en) * 2018-09-26 2019-01-15 西安空间无线电技术研究所 A kind of satellite-borne microwave radiometer antenna error bearing calibration
CN110617889A (en) * 2019-08-28 2019-12-27 西安空间无线电技术研究所 High-stability testing method applied to synthetic aperture microwave radiometer
CN110716185A (en) * 2019-10-22 2020-01-21 国家卫星气象中心 An on-orbit radiation correction method for spaceborne microwave radiometers
CN111538000A (en) * 2020-03-30 2020-08-14 西南电子技术研究所（中国电子科技集团公司第十研究所） Brightness Temperature Inversion Imaging Method for Uniform Circular Array Synthetic Aperture Radiometer
CN111929651A (en) * 2020-07-30 2020-11-13 西安空间无线电技术研究所 Method for calibrating phase residual error of synthetic aperture radiometer system
CN112285659A (en) * 2020-07-30 2021-01-29 西安空间无线电技术研究所 A method for on-orbit update of brightness temperature reconstruction matrix based on synthetic aperture radiometer
CN113218510A (en) * 2021-05-13 2021-08-06 上海航天测控通信研究所 Feed source mouth surface external calibration device of satellite-borne one-dimensional synthetic aperture radiometer
CN113484917A (en) * 2021-06-30 2021-10-08 中国电波传播研究所（中国电子科技集团公司第二十二研究所） Automatic calibration method for microwave radiometer

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
JIAKUN WANG 等: "In-Orbit External Calibration Method for Synthetic Aperture Radiometer", 《 IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING》, vol. 14, pages 3763, XP011849214, DOI: 10.1109/IGARSS.2019.8898941 *
吴琼, 吴季: "一种一维综合孔径微波辐射计的定标方法", 遥感技术与应用, no. 03, pages 53 - 56 *
赵晓雪;张晓辉;: "基于神经网络控制的星载微波辐射计", 微计算机信息, no. 13, pages 66 - 68 *
韩东浩: "旋转圆环阵列综合孔径微波辐射计系统研究", 《中国博士学位论文全文数据库 (信息科技辑)》, no. 09, pages 140 - 62 *

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115932843A (en) * 2022-10-28 2023-04-07 西安空间无线电技术研究所 Push-broom radiometer digital beam synthesis method based on multilayer perception network
CN117196418A (en) * 2023-11-08 2023-12-08 江西师范大学 An artificial intelligence-based reading teaching quality assessment method and system
CN117196418B (en) * 2023-11-08 2024-02-02 江西师范大学 Reading teaching quality assessment method and system based on artificial intelligence

Also Published As

Publication number Publication date
CN114046888B (en) 2024-02-20

Similar Documents

Publication Publication Date Title
CN101975947B (en) Two-dimensional radial imaging method for mirror image synthetic aperture
CN118050692B (en) A fast automatic calibration method for phased array based on residual neural network
CN114626987B (en) A physics-based electromagnetic inverse scattering imaging method for deep expansion networks
CN111190151B (en) System parameter design and transmit power optimization method for multimodal small satellite SAR in scanning mode
CN101625408A (en) Method for rarefying and optimizing energy transducer array of three-dimensional imaging sonar system
CN114046888A (en) Beam synthesis push-broom radiometer calibration method based on convolutional neural network
CN114895235B (en) A DOA estimation method for coprime arrays based on pseudo-snapshot increments in the presence of impulse noise
CN106597441A (en) Multi-target ISAR imaging task-oriented MIMO radar waveform optimal design method
CN118915064A (en) Sparse ISAR translational compensation and imaging integrated method based on super network
CN118971922A (en) Calibration method, device, system and storage medium for radio astronomy phased array feed
CN118330551A (en) Method, device and system for estimating angle of arrival based on reconfigurable intelligent super-surface assistance
CN117706553A (en) ISAR imaging method based on sparse Bayesian depth expansion network
CN117214893A (en) ISAR imaging method based on data model hybrid driven Bayesian expansion network
Korotetskiy et al. Phased array antenna calibration with probe positioning errors [Measurements Corner]
CN116148758A (en) Method, system and medium for on-orbit beam pointing prediction of radar antenna
CN119881978B (en) Joint inversion method of ocean surge height and wind wave height based on multimodal data fusion of spaceborne GNSS-R
CN106680809A (en) Self-focusing compressed sensing imaging method of through-wall radar
CN118688796B (en) Deep learning SAR-based one-dimensional sea wave spectrum inversion method, device and medium
CN115390060B (en) Target two-dimensional angle positioning method, device, radar and storage medium
CN109061645B (en) Virtual correlation sparse imaging method and system for synthetic aperture radiometer
CN119846543A (en) Direction of arrival estimation method, radar system and vehicle
CN115932843B (en) Push-broom radiometer digital wave beam synthesis method based on multilayer perception network
Chuanhao et al. A Multipath Signals Recognition Method based on Deep Neural Network
CN119416402B (en) Sparse antenna array synthesis method based on neural network intelligent solution
CN115902892B (en) A forward-looking imaging method for airborne radar based on alternating direction multiplier network

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant